Creating images with ChatGPT is done by describing what you want to see in the text field. ChatGPT then generates an image based on your description within seconds. The feature works in the web version, desktop app and mobile apps.<\/p>\n
You don't need any graphic knowledge. All that matters is a clear description. The more specific your prompt, the better the result will match what you have in mind. All users have access to the image function, even with a free account.<\/p>\n
ChatGPT's image function converts text to images via an underlying image model. Since late 2025, that is GPT Image 1.5, the successor to DALL-E 3. This model generates images up to four times faster than the previous version and follows instructions more accurately.<\/p>\n
Technically, ChatGPT does not generate existing images, but creates a new image based on patterns the model has learned during training. Each image is unique, even if you re-enter the same prompt. The model combines your description with learned visual patterns and produces a result that matches your instruction as closely as possible.<\/p>\n
GPT Image 1.5 is better than its predecessor at handling text in images, maintaining consistent styles and performing targeted edits. The model can also modify existing images without changing the rest of the image. Who Know how ChatGPT works<\/a>, recognises the principle: the system predicts the most likely visual outcome based on your input.<\/p>\n Creating an image in ChatGPT starts by opening a new conversation. Log in to chatgpt.com or open the app, and type a description of what you want to see. ChatGPT automatically recognises that you want an image and generates the image.<\/p>\n Write a description that is as specific as possible. Name the subject, style, colours, composition and atmosphere. \u201cA modern city at sunset with glass buildings and a pink sky\u201d will yield a more useful result than \u201ca beautiful picture of a city\u201d.<\/p>\n After generation, the image will appear in your conversation. You can download the image directly, or ask ChatGPT for adjustments. \u201cMake the sky darker\u201d or \u201cAdd a person in the foreground\u201d are examples of follow-up prompts that further refine the image. By refining step by step, you get closer to the result you had in mind.<\/p>\n On the mobile app, tap \u201cCreate an image\u201d in the prompt bar, which appears as a separate function. On desktop, just type your description in the text field. In both cases, you don't need to select a separate template; ChatGPT automatically turns on the image function as soon as it recognises that you want an image.<\/p>\n The quality of your prompt determines the quality of your image. ChatGPT makes the same choices when generating images as it does with text: it fills in missing details itself based on probability. The more context you provide, the less the model has to guess.<\/p>\n Name five elements in your prompt: the subject (what is the focus), the style (photorealistic, illustration, cartoon, oil painting), the composition (close-up, bird's eye view, symmetrical), the lighting (natural light, studio lighting, dramatic contrast) and the mood (warm, matter-of-fact, playful). Not every prompt needs to include all five, but the more you name, the more predictable the result.<\/p>\n Avoid vague adjectives such as \u201cbeautiful\u201d or \u201cprofessional\u201d. Those terms are subjective and give the model little guidance. Instead, describe concretely what you see: \u201ca white background with soft shadow\u201d is more useful than \u201ca professional look\u201d.<\/p>\n Work iteratively. Generate a first version, assess what is right and wrong, and refine your prompt based on the result. This process from framework to output to control is the same approach used in the ChatGPT course<\/a> of LearnLLM is central to working with text. With images, it works no differently.<\/p>\n Give your prompt in English if you find that results in Dutch are less accurate. The model is trained on predominantly English-language data, so English descriptions are sometimes interpreted more accurately. This is especially true for specific style descriptions and technical terms.<\/p>\n ChatGPT can edit existing images in addition to creating images. Upload an image via the paperclip icon and provide an instruction for the adjustment. The model adjusts only the requested part and leaves the rest of the image intact, including lighting, composition and facial features.<\/p>\n This feature is called \u201cinpainting\u201d and works by selecting a specific area in the image and re-generating it based on your instruction. For example, ask to change the background, add or remove an object, adjust clothing or put text on a product.<\/p>\n Please note that ChatGPT does not work with your original file, but generates a new version based on your instructions. For product photos or official imagery, this may produce subtle differences. Always check the result carefully before publishing.<\/p>\n Want to know more about uploading images to ChatGPT for analysis or editing? Read our article on how to upload an image in ChatGPT<\/a>.<\/p>\n Yes, you can create free images with ChatGPT. The free account provides access to the image function, but with a limited number of generations per day. Currently, the free version allows you to create 2-3 images per 24 hours. That limit works like a rolling window: if you create your first image at 10:00 am, that spot will become available again at 10:00 am the next day.<\/p>\n With ChatGPT Plus (about $20 per month), you get significantly more space: about 50 images per 3 hours. For most professionals, this is more than enough, even with more intensive use. ChatGPT Pro ($200 per month) offers unlimited image generation and is intended for users who need large amounts of images daily.<\/p>\n The free version is suitable for exploring the feature and assessing whether the quality suits your work. Structural professional use requires a Plus subscription. A full overview of subscriptions and their features can be found in our article on What ChatGPT costs<\/a>.<\/p>\n The number of images you can create with ChatGPT depends on your subscription. OpenAI works with rolling windows instead of fixed daily limits, meaning your limit does not reset at midnight but shifts continuously based on your usage.<\/p>\n The free account allows 2 to 3 images per 24 hours. ChatGPT Plus offers about 50 images per rolling 3-hour window, which amounts to about 200 per day if used properly timed. ChatGPT Pro offers unlimited generation, although extreme usage may still be restricted to prevent abuse.<\/p>\n Every generation attempt counts towards your limit, even if the result is not usable. Therefore, prepare your prompts well and avoid resending a prompt while ChatGPT is still generating. That costs an extra attempt with no result. Complex prompts can take up to 2 minutes of processing time; wait for the result before making another attempt.<\/p>\n During peak hours, OpenAI may temporarily lower the limits. There is no dashboard that shows how much of your quota you have already used, so you only notice the limit when you get the error message.<\/p>\n ChatGPT has image creation limitations that you should be aware of before using it for professional work. Results are not always error-free, and certain types of images are deliberately excluded.<\/p>\n OpenAI does not allow you to generate images of existing, recognisable people. If you ask for a portrait of a well-known public figure, ChatGPT will refuse the request. This policy is meant to prevent abuse and deepfakes. Generating images in the style of a specific living artist is also blocked to protect the intellectual property of creators.<\/p>\n Realistic human faces remain a concern. ChatGPT can generate convincing portraits, but errors frequently occur in details such as hands, fingers and teeth. Always check generated images with people carefully. Alternatives are drawn figures, silhouettes or abstract renderings.<\/p>\n Text in images is greatly improved with GPT Image 1.5, but not yet error-free. It works well for short words and sentences, but spelling errors or distortions can occur with longer text or specific fonts. Always check text in generated images letter by letter.<\/p>\n Finally, ChatGPT generates images, it does not take pictures. Each image is a new creation based on patterns, not a reproduction of reality. Do not use generated images as a replacement for real product photos on an online store, and be transparent towards your audience when deploying AI-generated images. Who the risks of AI<\/a> knows, knows that transparency builds trust.<\/p>\n ChatGPT images can be used for a variety of professional tasks where quick visual material is needed. The image function is not a replacement for a graphic designer, but it can speed up much of the visual preliminary work.<\/p>\n Social media content is one of the most common applications. Generate a visual to accompany a LinkedIn post, create variations for Instagram Stories or create a banner for a newsletter, without opening a separate design programme. In your prompt, specify the desired format and style for best results.<\/p>\n For presentations and reports, you can quickly have illustrations, concept visualisations or schematic views created. Upload an outline or describe a concept, and ChatGPT generates a visual that you can use as a starting point or as a final image in an internal presentation.<\/p>\n Product visualisations are useful to explore prototypes, packaging concepts or mood images before investing in a photo shoot. Use them as internal communication or briefing for a designer, not as a replacement for real product photos on your website.<\/p>\n It also offers value for educational content, infographics and blog illustrations. The combination of speed and flexibility makes it possible to create images that exactly fit the message, without depending on stock photos.<\/p>\n Want to know more about all the possibilities of ChatGPT for your daily work? Read our article on What you can do with ChatGPT<\/a>.<\/p>\nHow to create an image in ChatGPT step by step?<\/h2>\n
Tips for better prompts when creating images<\/h2>\n
Editing images in ChatGPT<\/h2>\n
Can you create free images with ChatGPT?<\/h2>\n
How many images can you create with ChatGPT?<\/h2>\n
Limitations of ChatGPT when creating images<\/h2>\n
What do you use ChatGPT images for in your work?<\/h2>\n
Comparison table image generation by subscription<\/h3>\n